Rule-Based Cyberbullying Detection on Social Media

PROJECT ABSTRACT:

Cyberbullying has become a growing concern in online social networking (OSN) platforms, leading to serious psychological and emotional consequences for users. To address this issue, we propose a Rule-Based Cyberbullying Detection System that identifies and prevents cyberbullying activities on social media. Unlike machine learning-based approaches, our system operates on a predefined set of rules established by the administrator. The system allows admins to maintain a customizable list of offensive words that serve as the basis for detecting cyberbullying in user-generated content.

The Online Social Networking (OSN) module forms the foundation of the platform, incorporating essential features such as user registration, authentication, public messaging, post sharing, user search, and friend requests. The cyberbullying detection mechanism actively scans messages and posts against the pre-configured list of offensive terms. If a user attempts to share content containing flagged words, the system immediately detects the violation and blocks the post or restricts the user accordingly.

This rule-based approach ensures real-time filtering of harmful content without the need for complex machine learning models, making it lightweight, efficient, and easy to manage. The system provides an effective and scalable solution for maintaining a safer online environment, reducing the spread of cyberbullying, and promoting positive digital interactions.

This project demonstrates the feasibility of rule-based content moderation in social media applications and lays the groundwork for future enhancements, such as sentiment analysis and dynamic rule expansion.

PROJECT OUTPUT VIDEO:

EXISTING SYSTEM:

In the existing system on computational studies of bullying have shown that natural language processing and machine learning are powerful tools to study bullying.
Cyberbullying detection can be formulated as a supervised learning problem. A classifier is first trained on a cyberbullying corpus labeled by humans, and the learned classifier is then used to recognize a bullying message.
Yin et.al proposed to combine BoW features, sentiment features and contextual features to train a support vector machine for online harassment detection.
Dinakar et.al utilized label specific features to extend the general features, where the label specific features are learned by Linear Discriminative Analysis. In addition, common sense knowledge was also applied.
Nahar et.al presented a weighted TF-IDF scheme via scaling bullying-like features by a factor of two. Besides content-based information, Maral et.al proposed to apply users’ information, such as gender and history messages, and context information as extra features
In the existing system, certain social networking platforms also incorporated content moderation teams to manually review flagged posts and take necessary actions, such as warning or banning users who violated platform guidelines. Many systems provided privacy settings that allowed users to control who could view their posts or interact with them, thereby reducing unwanted interactions.
Overall, the existing system established foundational content moderation methods, focusing on user reporting mechanisms and manual intervention to ensure safer digital interactions.

DISADVANTAGES OF EXISTING SYSTEM:

Hard to Describe: The first and also critical step is the numerical representation learning for text messages. Secondly, cyberbullying is hard to describe and judge from a third view due to its intrinsic ambiguities. Thirdly, due to protection of Internet users and privacy issues, only a small portion of messages are left on the Internet and most bullying posts are deleted.
Lacks of Context Awareness – The existing system filtering systems were unable to understand the context of a message. Words with multiple meanings could be misinterpreted, leading to either unnecessary censorship or failure to detect actual cyberbullying.
Dependency on Manual Moderation – The system heavily relied on human moderators to find the cyberbullying word and then review flagged content, which made the process slow, labor-intensive, and impractical for handling large-scale social media platforms with millions of users.
Inefficiency in Real-Time Monitoring – The manual reporting and review process caused delays in taking action against cyberbullying incidents. Harmful content could spread widely before moderators could intervene.
Inability to Handle Evolving Language – Cyberbullies often use creative ways to bypass detection, such as using slang, abbreviations, special characters, or code words. The static rule-based approach struggled to adapt to these evolving tactics.
Limited User Protection Mechanisms – The earlier systems primarily focused on detecting offensive words but lacked features such as automated warnings, temporary restrictions, or advanced behavior tracking to prevent repeat offenders.
Scalability Issues – As the volume of online interactions increased, maintaining a manually curated blacklist of offensive words became difficult, making the system less effective for large-scale social media platforms.
Over-Reliance on User Reporting – Many platforms depended on users to report cyberbullying incidents, leading to delayed action and potential bias, as not all victims or bystanders actively reported harmful behavior.

PROPOSED SYSTEM:

The Rule-Based Cyberbullying Detection System is designed to enhance content moderation on social media by automatically detecting and preventing the use of offensive language in user-generated posts, messages, and comments. The system operates based on a predefined set of rules and keywords, managed by the administrator, to identify cyberbullying-related content and take appropriate action.
The system is integrated within an Online Social Networking (OSN) module, which includes core functionalities such as user registration, authentication, messaging, post sharing, friend requests, and profile searches. Users can interact within the platform, share content, and communicate, while the system continuously monitors text-based interactions for any flagged words or phrases.
A centralized database of cyberbullying keywords is maintained by the administrator. When a user attempts to post or send a message containing any of flagged words, the system automatically detects the content. Additionally, the system can restrict user actions based on repeated violations, ensuring a controlled and safer environment for users.
The proposed system utilizes real-time text scanning and rule-based filtering techniques, ensuring immediate detection of offensive content. The rules can be customized and updated by the administrator to adapt to emerging cyberbullying trends and maintain the system’s effectiveness.
This non-AI-driven approach ensures that the system remains lightweight, easily manageable, and efficient for implementation in social networking platforms without requiring complex machine learning models or extensive computational resources. The system aims to establish a structured content moderation framework, ensuring that social media interactions remain free from harmful or abusive language.

ADVANTAGES OF PROPOSED SYSTEM:

Real-Time Cyberbullying Detection – The proposed system instantly detects the cyberbullying words automatically, don’t require to find manually, ensuring immediate prevention of harmful interactions.
Lightweight and Efficient – Since the system is rule-based, it does not require complex machine learning models or high computational power, making it faster and more resource-efficient.
User Safety and Moderation – By automatically blocking harmful content and restricting users who violate the rules, the system ensures a safer online environment for all users.
No Training Data Required – Unlike AI-based models, which require large datasets for training, the system works solely on predefined rules, reducing the need for continuous learning and retraining.
Scalable for Social Media Platforms – The system can be integrated into various online social networks, making it a versatile solution for different types of digital communities.
Simple Implementation and Maintenance – With a straightforward rule-based mechanism, the system is easy to deploy and maintain without requiring advanced technical expertise.
Minimization of Manual Moderation – Reduces the need for human moderators by automating the detection and filtering of offensive content, making moderation more efficient.
Prevention of False Reports – Unlike user-reporting-based systems, this approach automatically enforces predefined rules, eliminating bias or misuse of the reporting feature.
Ensures Platform Integrity – By actively moderating and filtering abusive language, the system helps maintain a positive and respectful online environment for all users.
Admin-Controlled Customization – Administrators can add, update, or modify the list of offensive words, making the system flexible and adaptable to evolving cyberbullying trends.
These advantages make the Rule-Based Cyberbullying Detection System an effective solution for preventing online harassment and ensuring safer digital interactions on social media platforms.

SYSTEM REQUIREMENTS:

HARDWARE REQUIREMENTS:

System : Pentium i3 Processor.
Hard Disk : 500 GB.
Monitor : 15’’ LED.
Input Devices : Keyboard, Mouse.
Ram : 4 GB.

SOFTWARE REQUIREMENTS:

Operating system : Windows 10/11.
Coding Language : JAVA.
Frontend : JSP, HTML, CSS, JavaScript.
JDK Version : JDK 23.0.1.
IDE Tool : Apache Netbeans IDE 24.
Tomcat Server Version : Apache Tomcat 9.0.84
Database : MYSQL.

Java Projects

Rule-Based Cyberbullying Detection on Social Media